AI Researchers

About Alt Inc

Alt Inc builds AI-first products and platforms that turn cutting-edge research into reliable, real-world impact. We value curiosity, strong fundamentals, and pragmatic experimentation—shipping learnings as well as models.

Role Overview

As an AI Researcher at Alt Inc (Bengaluru), you will explore, prototype, and validate machine learning approaches across areas such as LLMs, multimodal learning, retrieval, and applied deep learning. This role is ideal for early-career researchers and engineers (0–5 years) who enjoy reading papers, running rigorous experiments, and translating results into scalable solutions.

What You'll Do

* Design and run experiments to evaluate model architectures, training strategies, and inference-time techniques * Fine-tune and adapt foundation models (LLMs/vision/multimodal) for product use-cases * Build and evaluate RAG pipelines: retrieval, reranking, chunking strategies, and grounding/faithfulness checks * Develop datasets and labeling strategies; define metrics and evaluation harnesses (offline + online) * Implement training/inference pipelines with attention to reproducibility, efficiency, and cost * Collaborate with engineering and product teams to translate research prototypes into production-ready components * Document findings clearly (reports, experiment logs, ablations) and communicate trade-offs

Required Skills

* Strong foundations in machine learning and deep learning (optimization, generalization, representation learning) * Proficiency in Python and modern ML tooling (e.g., PyTorch/JAX, Hugging Face ecosystem) * Hands-on experience running experiments: data preprocessing, training loops, hyperparameter tuning, ablations * Comfort with scientific thinking: forming hypotheses, designing controlled experiments, interpreting results * Solid software engineering basics: clean code, version control (Git), debugging, and testing mindset * Good communication skills—able to explain results and uncertainty to technical and non-technical stakeholders

Good to Have

* Experience with LLM fine-tuning (SFT, preference optimization), prompt engineering, and evaluation * Experience with retrieval systems (BM25, dense retrieval, vector databases) and reranking * Knowledge of distributed training/inference, GPU performance basics, and model optimization (quantization, distillation) * Familiarity with MLOps practices (experiment tracking, model registry, CI/CD for ML) * Publications, open-source contributions, or strong project portfolio demonstrating research rigor

What Success Looks Like

* You can independently take a research question from idea → experiment plan → results → recommendation * You build reliable evaluation setups that catch regressions and measure real improvements * You deliver prototypes that are easy for engineering teams to productionize * You communicate clearly: what worked, what didn't, and what to try next

Location & Experience

* Location: Bengaluru, India * Experience: 0–5 years